|Publication number||US20080304712 A1|
|Application number||US 12/181,495|
|Publication date||Dec 11, 2008|
|Filing date||Jul 29, 2008|
|Priority date||Jun 5, 2001|
|Also published as||CA2448761A1, CN1516564A, CN100415164C, DE60232794D1, EP1397667A2, EP1397667B1, US7613504, US7890158, US9061899, US20020183624, US20030078504, US20080297788, US20130021596, US20130027685, WO2002099393A2, WO2002099393A3|
|Publication number||12181495, 181495, US 2008/0304712 A1, US 2008/304712 A1, US 20080304712 A1, US 20080304712A1, US 2008304712 A1, US 2008304712A1, US-A1-20080304712, US-A1-2008304712, US2008/0304712A1, US2008/304712A1, US20080304712 A1, US20080304712A1, US2008304712 A1, US2008304712A1|
|Inventors||Robert K. Rowe, Stephen P. Corcoran, Shonn P. Hendee|
|Original Assignee||Lumidigm, Inc.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (7), Referenced by (6), Classifications (24), Legal Events (1)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application is a division of U.S. patent application Ser. No. 09/874,740, filed Jun. 5, 2001, entitled “Apparatus and Method of Biometric Determination Using Specialized Optical Spectroscopy Systems” (“the parent application”) which is incorporated herein by reference for all purposes. The parent application is related to U.S. patent application Ser. No. 09/832,534, filed Apr. 11, 2001, entitled “Apparatus and Method of Biometric Identification or Verification of Individuals using Optical Spectroscopy”, which is a continuation-in-part of U.S. patent application Ser. No. 09/415,594, filed Oct. 8, 1999, entitled “Apparatus and Method for Identification of Individuals by Near-Infrared Spectrum”; which is related to U.S. patent application Ser. No. 09/174,812, filed Oct. 19, 1998, entitled “Method for Non-Invasive Analyte Measurement with Improved Optical Interface”; and U.S. patent application Ser. No. 08/871,366, filed Jun. 9, 1997, entitled “Diffuse Reflectance Monitoring Apparatus”, all assigned to the same assignee as the present application, and the disclosures of which are incorporated herein by reference.
This present invention relates generally to methods and systems for performing biometric determinations of individuals utilizing optical spectra of tissue. More specifically, the invention relates to methods and systems for determining or verifying identity, determining or verifying age, determining or verifying sex, and determining or verifying liveness and authenticity of the sample being measured. The present invention discloses methods and systems for gathering optical information about the tissue using a combination of wavelengths and source-detector separations. The present invention discloses a family of compact, special-purpose optical sensors operating in the near-ultraviolet, visible, and near-infrared spectral regions that are suitable for a variety of biometric determination tasks. The sensors can be used in stand-alone, dedicated applications or can be incorporated in a variety of personal devices such as cellular telephones, personal digital assistants, wrist watches, or electronic fobs to provide personal biometric security to protect access to a variety of protected property.
Biometric determination is generally defined as the process of measuring and using one or more physical or behavioral features or attributes to gain information about identity, age, or sex of a person, animal, or other biological entity. As well, in order to ensure security, the biometric determination task may include further tasks that ensure that the sample being measured is authentic and being measured on a living being. This latter test is referred to as a determination of liveness.
There are two common modes in which biometric determinations of identity occur: one-to-many (identification) and one-to-one (verification). One-to-many identification attempts to answer the question of, “do I know you?” The biometric measurement device collects a set of biometric data and from this information alone it assesses whether the person is a previously seen (“authorized”) individual. Systems that perform the one-to-many identification task, such as the FBI's Automatic Fingerprint Identification System (AFIS), are generally very expensive ($10 million or more) and require many minutes to detect a match between an unknown sample and a large database containing hundreds of thousands or millions of entries. The one-to-one mode of biometric analysis answers the question of, “are you who you say you are?” This mode is used in cases where an individual makes a claim of identity using a user name, a personal identification number (PIN) or other code, a magnetic card, or other means, and the device collects a set of biometric data which it uses to confirm the identity of the person.
Although in general the one-to-many identification task is more difficult than one-to-one, the two tasks become the same as the number of recognized or authorized users for a given biometric device decreases to just a single individual. Situations in which a biometric identification task has only a small number of entries in the authorization database are quite common. For example, biometric access to a residence, to a personal automobile, to a personal computer, to a cellular telephone, and to other such personal devices typically require an authorization database of just a few people.
Biometric identification and verification is useful in many applications. Examples include verifying identity prior to activating machinery or gaining entry to a secure area. Another example would be identification of an individual for matching that individual to records on file for that individual, such as for matching hospital patient records especially when the individual's identity is unknown. Biometric identification is also useful to match police records at the time a suspect is apprehended, but true identity of the suspect is not known. Additional uses of biometric identification or verification include automotive keyless start and entry applications, secure computer and network access applications, automated financial transaction applications, authorized handgun use applications, and time-and-attendance applications. In general, protected property will be the term used to describe all of the goods, places, services, and information that may require biometric authorization to access.
Current methods for biometric identification are manifold, but some of the most common techniques include fingerprint pattern matching, facial recognition, hand geometry, iris scanning, and voice recognition. Each of these technologies addresses the need for biometric identification to some extent. However, due to cost, performance, or other issues, each of the existing methods has advantages and disadvantages relative to the other technologies.
There are currently many personal electronic devices that are used to gain access to protected property but that do not include any biometric capability. For example, electronic fobs are commonly used to gain entry to automobiles and to activate commercial and residential alarm systems. Wristwatches such as the Swatch Access models can be used to purchase and download codes that allow easy entry to ski areas and other for-pay recreational sites. A wristwatch being sold by Xyloc permits access to computers, printers, networks, or other properly equipped hardware and systems when the watch is in the vicinity of the protected system. A small electronic device known as an iButton sold by Dallas Semiconductor can be put into a ring, key fob, wallet, watch, metal card or badge, that a person can carry and use to gain access to properly equipped doors and other protected systems. However, an unauthorized user can gain access to any of the property protected by these systems by simply obtaining a device from an authorized user. These devices do not have the capability to distinguish between authorized and unauthorized users and will work for anyone who possesses them. This deficiency represents a major security concern.
In U.S. Pat. No. 6,041,410, Hsu et al. disclose a personal identification fob that employs fingerprint data. This system is specified to contain memory to hold the fingerprint image, an image correlater, a communication means employing a cyclic redundancy code, and a “door” that is controlled by the biometric system and allows access to protected property. Hsu et al. generalize “door” as a means to access protected property including a building, a room, an automobile, and a financial account. The method disclosed relates to a door that protects property and its interaction with the fob, including a “wake-up” message and a series of steps to collect the biometric data and compare it with reference data, determining a match, and then actuating the device to provide access through the door.
One company that currently sells a personal identification unit is affinitex, a division of AiT, and the product name is VeriMe. Because of the size of the fingerprint reader incorporated in the VeriMe product as well as the batteries and control electronics, the unit is relatively large and is intended to be hung around the neck like a pendant. In contrast, a long-standing desire of many in the biometric community is a biometric technology that can be discretely incorporated in a piece of jewelry such as a wristwatch (for example, see Biometrics; Advanced Identity Verification, Julian Ashbourn, Springer, 2000, pp. 63-4).
There are a number of known biometric products and technologies that rely on optical images of various tissue sites to perform a biometric determination. For example, in U.S. Pat. No. 4,537,484, Fowler, et al. describe an apparatus for collecting a fingerprint image using optical techniques. In U.S. Pat. No. 6,175,407, Sartor describes an apparatus for collecting a palm image using optical techniques. In U.S. Pat. No. 5,291,560, Daugman describes a method for collecting and processing an optical image of the iris. In U.S. Pat. No. 5,793,881, Stiver et al. describe a system and method for collecting an image of the subcutaneous structure of the hand using an imaging methodology. However, all of these technologies generate and use images of the tissue as the basis for a biometric determination. The use of imaging generally requires high-quality expensive optical systems and an imaged region that is of sufficient size to capture the necessary biometric detail. If the imaged region is made too small, the biometric performance of these imaging systems degrade. For this reason, contact imaging systems such as fingerprint and palm readers require a relatively large, smooth, accessible surface, limiting the range and form of products in which such systems can be incorporated. Finally, because the determination of a match between enrolled images and the test images is dependent on the orientation of the two images, such biometric systems have to correct for these positional effects. For this reason, biometric systems that rely on imaging techniques require a significant computational power and a sophisticated algorithm to correct for image displacements, rotations and distortions, which leads to increased system cost and increased time required for user authentication.
As an alternative to imaging techniques, the use of spectral information for biometric determinations is disclosed in U.S. patent application Ser. No. 09/832,534, filed Apr. 11, 2001, entitled “Apparatus and Method of Biometric Identification or Verification of Individuals using Optical Spectroscopy”, which is a continuation-in-part of U.S. patent application Ser. No. 09/415,594, filed Oct. 8, 1999, entitled “Apparatus and Method for Identification of Individuals by Near-Infrared Spectrum”. The equipment used to perform the measurements disclosed in these applications was based on relatively large and expensive multi-purpose laboratory-grade commercial spectrometers. The family of techniques disclosed in these applications is referred to as spectral biometrics. The disclosures of these applications are incorporated herein by reference.
It is well known that tissue spectra are generally affected by both the absorption and scattering properties of the tissue. For many spectral measurement applications the portion of the measured spectra that represent the absorption characteristics of the tissue are more important for the measurement rather than the effects due to scatter. One technique for separating the two effects is known as radially resolved diffuse reflectance spectroscopy, which is based on collecting multiple measurements with different source-detector separation distances. This collection of data provides enough information to estimate and separate effects due to scatter and absorption (see Nichols, et al., Design and Testing of a White-Light, Steady-State Diffuse Reflectance Spectrometer for Determination of Optical Properties of Highly Scattering Systems, Applied Optics, Jan. 1, 1997, 36(1), pp 93-104). Although the use of multiple source-detector separations is a well-known technique for analyte measurements in biological samples, the use of similar measurement configurations for spectral biometric determinations has not been previously disclosed.
There is a need for an inexpensive, rugged and small spectrometer to perform spectral biometric determinations. One method that can be used to construct such spectrometers is based on using multiple discrete light sources such as light emitting diodes (LEDs), laser diodes, vertical cavity surface emitting lasers (VCSELs), and narrow band optical filters coupled to a broad-band optical source such as an incandescent bulb or blackbody emitter, operating at different wavelengths to illuminate and measure the optical properties of the sample at each of these wavelengths. These types of spectrometers are known and used for collecting spectrometric information for many applications. For example, in U.S. Pat. No. 3,910,701, Henderson et al. disclose a spectrometer that incorporates a plurality of LED sources for measuring a variety of biological samples. In U.S. Pat. No. 4,857,735, Noller discloses a spectrometer using one or more LEDs to measure solution samples. In U.S. Pat. No. 5,257,086, Fately et al. disclose an optical spectrometer having a multi-LED light source incorporating Hadamard or Fourier frequency encoding methods. However, there is a need for a small, rugged, and inexpensive spectrometer with designs that are optimal for biometric determinations.
As part of the biometric determination task, there is a need for ensuring that the sample being used for the biometric determination is alive. For example, U.S. Pat. No. 5,719,950 to Osten et al. disclose a method and system to combine a biometric-specific measurement such as fingerprints, palm prints, voice prints, etc with a separate measurement of a non-specific biometric parameter such as skin temperature, pulse, electrocardiogram or tissue spectral features to ensure the liveness of the sample.
In addition to performing a biometric identification or verification and ensuring that the sample being measured is living tissue, there may also exist a need to determine an estimate of the age, sex, and other demographic characteristics of the person under test as part of the biometric determination task. For example, the U.S. Federal Trade Commission recently established a commission to examine the issue of remotely determining age of a person who is attempting to access a web site in order to block access by children to inappropriate sites. The Commission on Online Child Protection (COPA) heard testimony on Jun. 9, 2000 that indicated that then-known biometric techniques could not be used to aid the determination of a person's age based on any known biometric features.
Detailed embodiments of the present invention are disclosed herein. However, it is to be understood that the disclosed embodiments are merely exemplary of the present invention, which may be embodied in various systems. Therefore, specific details disclosed herein are not to be interpreted as limiting, but rather as a basis for the claims and as a representative basis for teaching one of skill in the art to variously practice the invention.
The present invention is based on Applicant's recognition that an accurate, precise and repeatable tissue spectra of an individual including selected wavelengths in the near ultraviolet range, visible range, very near infrared range or near infrared range and combinations of selected wavelengths from these ranges contains spectral features and combinations of spectral features which are unique to that individual. The spectral range over which biometric determinations have been demonstrated span wavelengths from 350 nm to 2500 nm, although it is likely that similar capabilities exist outside of this range. The choice of which measurement wavelengths to use is driven in part by the availability and cost of suitable illumination sources and detectors. In the case of the discrete light sources disclosed in this application, the most common and least expensive optical components work with light in the wavelength region from 350-1000 nm. Such a system can be constructed from silicon detector material and readily available LEDs, laser diodes, VCSELs, or optical filters coupled to a bulb. However, other detectors and other light sources could also be used as alternative components or to span a greater spectral range or a different spectral range.
The present method and apparatus also provides for biometric determination of whether a sample being measured is living tissue, known as a “liveness” determination. Further, the present system maintains high system security because the biometric device ensures that the exact sample it is operating on is real and alive, in addition to matching the properties of the enrolled data. Thus, an accurate determination of liveness precludes the use of simulated body parts and/or parts that have been removed from authenticated individuals. It has been found that the spectroscopic signature of living tissue is substantially different than most other media (including dead tissue), and thus liveness determination is an integral part of the present biometric device and method.
The present method and apparatus can also be used to estimate or verify the age of a person undergoing the biometric measurement. Further, the present method and apparatus can be used to estimate or verify the sex of the person undergoing the biometric measurement.
A variety of embodiments are disclosed herein for a sensor apparatus that can obtain tissue spectra that can be utilized for biometric identification determinations, liveness determinations, age determinations and sex determinations. These embodiments of the present invention are amenable to miniaturization and ruggedization for incorporation in a variety of systems. Such fixed-installation applications include, but are not limited to, physical entry assurance to workplaces, homes, hotels, secure industrial areas and other controlled sites; time and attendance monitoring; automotive applications such as keyless entry, keyless start, automotive personality setting, and mobile internet access; personal computer and network security; secure health record access; automated financial transactions; and authorized handgun use.
In addition to the fixed-system applications, the apparatus and methods disclosed in this application can be used in small personal biometric packages such as smart cards, electronic fobs or wristwatches that the user/owner/wearer can carry with them and provide biometrically-assured authorization to a variety of devices and systems. Such personal biometric systems act as keys that provide access to protected property only if activated by the authorized individual, and reduces or eliminates entry to the protected system by unauthorized people. Thus, the personal biometric devices of the present invention become a type of smart key that allow access to any system that interfaces to such device and for which the holder is authorized to access. Such systems and uses can include, but are not limited to, personal computers, network access devices, doors in office buildings and private residences, time-and-attendance systems, automobiles, security equipment, automated financial transactions, cellular telephones, toll booths, electronic vending machine transactions and pay-per-entry events such as movies, etc.
Alternatively, as personal electronics such as personal digital assistants (PDA) and cellular phones become integrated in a variety of wireless applications, the present invention provides a means to confirm the identity of the person using the device. This can be important when wireless applications such as mobile commerce use such devices to authorize monetary transfers or make purchases, while also allowing access to medical records and act as an electronic key for homes, offices and automobiles. By providing an integrated, compact, rugged, secure biometric system that can be used to confirm the identity of a person attempting to use the PDA to access protected property, the present invention provides a capability that is applicable in many everyday life situations.
In addition to the application of the spectral biometric sensor as a single biometric, the sensor and identification methods disclosed in this application can also be used in conjunction with other biometric techniques within a system to either increase the accuracy of the system or increase the robustness of the system. In cases where greater system security is required, the spectral biometric technique may be combined with one or more other biometric methods and the results can be combined to ensure a person's identity. Alternatively, the disclosed systems and methods can be combined with other biometric techniques to offer more than one method to identify a person in case one method is disabled due to system failure or other reason, ensuring a more robust system performance overall.
One system for performing biometric determinations includes: an optical sensor head consisting of one or more monochromatic illumination sources, one or more detectors, and an optical sampler, all arranged such that there exists a plurality of source-detector spacings or a plurality of different monochromatic wavelengths, or both; a light-source encoding system; a microprocessor with an input and output device; a database including selected tissue spectral data for authorized persons or a collection of spectral data for individuals against which unknown individual's would be checked; and a program running in the microprocessor for discriminating between a target individual's spectral data and the authorized spectral data or collection of spectra database containing spectra for a group of individuals. The program can include software for performing an analysis for liveness determination, age determination, and sex determination based on the measured spectral data.
These and various other advantages and features of novelty which characterize the present invention are pointed out with particularity in the claims annexed hereto and forming a part hereof. However, for a better understanding of the invention, its advantages, and the object obtained by its use, reference should be made to the drawings which form a further part hereof, and to the accompanying descriptive matter in which there are illustrated and described preferred embodiments of the present invention.
In the drawings, in which like reference numerals indicate corresponding parts or elements of preferred embodiments of the present invention throughout the several views:
Detailed embodiments of the present invention are disclosed herein. However, it is to be understood that the disclosed embodiments are merely exemplary of the present invention which may be embodied in various systems. Therefore, specific details disclosed herein are not to be interpreted as limiting, but rather as a basis for the claims and as a representative basis for teaching one of skill in the art to variously practice the invention.
The present invention is based on Applicant's recognition that an accurate, precise and repeatable tissue spectrum of an individual in the near ultraviolet range, visible range, very near infrared, or near infrared spectral range and combinations of these ranges contains spectral features and combinations of spectral features which are unique to that individual. The present invention is further based on a recognition that proper analysis, utilizing discriminant analysis techniques, can identify these unique features or combinations, which are not readily apparent in visual analysis of a spectral output, so that an individual's identity may be determined by comparison of tissue spectral data taken at the time of use and compared to stored tissue spectral data from prior measurement.
In addition, the tissue spectrum has been found to not only contain information that is unique to an individual, but also contains numerous features and combinations of features that indicate whether such spectral samples were taken while the sample was alive or not. The physiological effects that give rise to spectral features that indicate the state of a sample (alive or dead) include but are not limited to blood perfusion, temperature, hydration status, glucose and other analyte levels, and overall state of tissue decay. Thus, the biometric identification and verification methods of the present invention can be also used in conjunction with, or separately from, the determination of the state of the liveness of the tissue. Tissue from other biological systems (organs, animals, etc.) has also been found to have spectral characteristics that are distinctly different from human skin due to differences in the tissue composition and form. Thus, the biometric identification methods of the present invention can be also used in conjunction with or separately from the determination of whether the sample is human skin or some other tissue. In addition, it has been found that tissue-like substances such as collagen gelatin, latex, water solutions, or others have spectral characteristics that are distinctly different than human tissue due to differences in composition and form. The biometric identification and verification methods of the present invention can thus be used with or separately from the determination whether the sample is actual tissue or some other substance.
While utilizing the present invention, it has also been found that other spectral features observed in the tissue spectrum relate to the age and sex of the person being measured. It is believed that these features are due in part to the differences in dermal thickness between young and old people and between males and females. Such changes in skin thickness and composition affect the optical characteristics of the tissue by affecting the scattering properties of the sample. These properties in turn impose distinct spectral shapes on the measured tissue spectra, which can be extracted and used by appropriate multivariate techniques to provide age and sex estimates.
Referring now to
The light sources 34 can be light emitting diodes (LEDs), laser diodes, vertical cavity surface emitting lasers (VCSELS), quartz tungsten halogen incandescent bulbs with optical pass-band filters with optical shutters, or a variety of other optical sources known in the art. The light sources 34 can each have the same wavelength characteristics or can be comprised of sources with different center wavelengths in the spectral range from about 350 nm to about 2500 nm. In general, the collection of light sources 34 can include some sources that have the same wavelengths as others and some sources that are different. In a preferred embodiment, the light sources 34 includes sets of LEDs, laser diodes, VCSELs, or other solid-state optoelectronic devices with differing wavelength characteristics that lie within the spectral range from about 350 nm to about 1100 nm.
The detector 36 can be a single element or it can be a one- or two-dimensional array of elements. The detector type and material is chosen to be appropriate to the source wavelengths and the measurement signal and timing requirements. These detectors can include PbS, PbSe, InSb, InGaAs, MCT, bolometers and micro-bolometer arrays. In a preferred embodiment where the light sources 34 are solid-state optoelectronic devices operating in the spectral range from about 350 nm to about 1100 nm, the preferred detector material is silicon.
The light sources 34 can be sequentially illuminated and extinguished to measure the tissue properties for each source by turning power to each of them on and off. Alternatively, multiple light sources 34 can be electronically modulated using encoding methods that are known to one knowledgeable in the art. These encoding patterns include Fourier intensity modulation, Hadamard modulation, random modulation, and other modulation methods.
In general, the detector 36 can be located in the center of the sensor head or it can be offset to one side of the sensor head 32 in order to provide for greater source-detector separation distances. The sensor head 32 can be other shapes including oval, square and rectangular. The sensor head 32 can also have a compound curvature on the optical surface to match the profile of the device in which it is mounted.
Light that reflects from the topmost layer of skin does not contain significant information about the deeper tissue properties. In fact, reflections from the top surface of tissue (known as “specular” or “shunted” light) are detrimental to most optical measurements. For this reason,
An alternative embodiment of a variable source-detector configuration is illustrated in
Alternatively, multiple source-detector distances can also be achieved by using more than one detector element as shown in
The use of multiple detector elements and multiple illumination sources can be extended to using a detector array as shown in
The detector(s) can be any material appropriate to the spectral region being detected. For light in the region from about 350 nm to about 1100 nm, a preferred detector material is silicon and can be implemented as a single-element device, a collection of discrete elements, or a 1-D or 2-D array, depending upon the system configuration and encoding method used. For light in the region from about 1.25 to about 2.5 .mu.m, a preferred detector material is InGaAs and can also be implemented as a single element, a collection of elements, or a 1-D or 2-D array. Additional detector materials and means of detection include InSb, Ge, MCT, PbS, PbSe, bolometers, and others known to one of ordinary skill in the art.
Once the light passing though the tissue is detected, the signals can be digitized and recorded by standard techniques. The recorded data can then be processed directly or converted into absorbance spectra or noised-scaled absorbance spectra as is known to one of ordinary skill in the art. The data can then be used for spectral identification or verification by the methods described in U.S. patent application Ser. No. 09/832,534, filed Apr. 11, 2001, entitled “Apparatus and Method of Biometric Identification or Verification of Individuals using Optical Spectroscopy”, and U.S. patent application Ser. No. 09/415,594, filed Oct. 8, 1999, entitled “Apparatus and Method for Identification of Individuals by Near-Infrared Spectrum”.
A small spectral biometric subassembly, such as those discussed above, can be embedded in a variety of systems and applications. The spectral biometric reader can be configured as a dedicated system that is connected to a PC or a network interface, an ATM, securing an entryway, or allowing access to a particular piece of electronics such as a cellular phone. In this mode, one or more people can be enrolled in the biometric system and use a particular reader to gain access to a particular function or area.
Alternatively, the spectral biometric system can configured as a personal biometric system that confirms the identity of the sole person authorized to use the device, and transmits this authorization to any properly equipped PC, ATM, entryway, or piece of electronics that requires access authorization. One advantage of this latter approach is that the personal biometric system can transmit an identifying code to the requesting unit and then use the biometric signal to confirm authorization, which implies that the system needs to perform a verification task rather than the more difficult identification task. Yet, from the user's perspective, the system recognizes the user without an explicit need to identify himself or herself. Thus, the system appears to operate in an identification mode, which is more convenient for the user.
An additional advantage of a personal biometric system is that if an unauthorized person is able to defeat the personal biometric system code for a particular biometric system-person combination, the personal biometric system can be reset or replaced to use a new identifying code and thus re-establish a secure biometric for the authorized person. This capability is in contrast to multi-person biometric systems that base their authorization solely on a biometric signature (spectral, as well as any of the other biometric techniques such as fingerprint, iris, facial, etc.). In this latter case, if an intruder is able to compromise the system by somehow imitating the signal from an authorized user, there is no capability to change the biometric code since it is based solely on a fixed physiological characteristic of a person.
A second embodiment of a personal spectral biometric system 110 is depicted in
In addition to the watch or fob, similar biometric capability can be built into other personal electronic devices. These devices include personal digital assistants (PDAs) and cellular telephones. In each case, the personal biometric system can provide user authorization to access both the device in which it is installed, as well as to provide authorization for mobile commerce (M-Commerce) or other wireless transactions that the device is capable of performing.
The compact sensors disclosed can also be put into firearms to prevent unauthorized usage. In particular, the biometric sensor could be placed in the handgrip of a weapon such as a handgun or other firearm to sense tissue properties while the gun is being held in a normal manner. A further capability of the apparatuses and methods disclosed in this application is the ability to identify people who are to be explicitly excluded from accessing protected property as well as determining those who are authorized to access the property. This capability will improve the biometric performance of the system with respect to those unauthorized people who are known to attempt to use the device, which could be particularly important in the case of a personal handgun. In particular, parents who own a biometrically enabled handgun can enroll themselves as authorized users and also can enroll their children as explicitly unauthorized users. In this way, parents could have further insurance that children who are known to be in the same household as a gun will not be able to use it.
It is also possible to use the explicit-denial capability of a biometric system in a fixed installation such as a home, place of business, or an automobile. For example, a biometric system installed at the entryway of a place of business can be used to admit authorized employees and temporary workers. If an employee is fired or the term of the temporary employee expires, then their enrollment data can be shifted from the authorized to the unauthorized database, and an explicit check is made to deny access to the former employee if he or she attempts to enter.
Because of the nature of optical spectroscopy, it is difficult to generate spectra of similar shape and absorbance characteristics without using similar material for the sample. For this reason, many common materials, such as latex and wax that are used to defeat other biometric systems such as fingerprint readers or hand geometry systems are ineffective tissue surrogates for a spectral biometric system. By performing a spectral comparison, most non-tissue samples will be rejected, resulting in a strong countermeasure capability against potential intruders.
Similarly, many of the spectral features that are present in the wavelength ranges disclosed by this invention are indicative of living tissue. These features include oxy- and deoxy-hemoglobin bands, temperature effects, intracellular hydration, and others. These effects contribute to the overall spectral signature of the sample being measured and ensure that a matching sample is one that is part of a living person and normally perfused. Thus, a good spectral comparison ensures the “liveness” of a sample and deters the use of dead or excised tissue as a means to circumvent the spectral biometric system.
In some applications, such as Internet access authorization, it may be useful to be able to verify the sex and/or age of the person using the spectral biometric system. Because of both age- and sex-specific difference in skin structure and composition, the optical spectra change in systematic and indicative ways such that the age and sex can be estimated using the biometric spectral data.
In practicing the present invention, the tissue spectral data is determined by measuring the light intensity received by the output sensor for the various light sources which give indications of the optical properties of the tissue at different wavelengths and/or at different source-detector separations. As is well known to one of ordinary skill in the art, the signal produced by the detector in response to the incident light levels can be converted into spectral data that can be recorded and used for subsequent analysis for enrollment or authorization of identity.
A laboratory experiment was performed to test and confirm the premise that discrete wavelength light sources could be used for biometric determination tasks and that further advantage could be gained by arranging the same sources with different source-detector spacings.
The optical sampler 106 included a sample head 120 which was capable of collecting tissue spectral data using two different source-detector spacings.
Twenty-two diabetic subjects participated in a study, which spanned a total duration of 16 weeks. Each person in the study was measured during two separate visits per week for each of the first 7 weeks of the study. There was then an 8-week gap, followed by one additional week of study where each person again was measured during two separate visits. During each measurement visit, multiple (5) optical samples were collected from the underside of their left forearm. Each optical sample consisted of 90 seconds of measurement time.
The optical samples collected by the sampler shown in
The calibration data were processed to produce generic data as described in U.S. Pat. No. 6,157,041, entitled “Methods and Apparatus for Tailoring Spectroscopic Calibration Models”. A PCA decomposition of these data was performed to generate 50 eigenvectors and scores. The scores were then analyzed to determine the 20 factors that had the largest values for the ratio of the between-person variation to the within-person variation for each set of scores.
The first two samples for each of the validation subject's data were averaged and used as the initial enrollment spectra. Each of the remaining validation spectra were taken in temporal sequence and subtracted from the enrollment spectrum. This spectral difference was then presented to the selected calibration factors and a Mahalanobis distance was calculated. If the Mahalanobis distance was below a certain threshold value, the validation spectrum was deemed valid, and a weighted sum of the validation spectrum (0.2) and the enrollment spectrum (0.8) was used to update the enrollment spectrum. This process was repeated for multiple threshold values. One of ordinary skill in the art will recognize that the Spectral F-Ratio could be used instead of or in conjunction with the Mahalanobis distance metric to perform the identity determinations. The intruder data was processed in a similar manner as the validation data using the same threshold values.
This analysis was applied to spectral data from inner ring illumination, from outer-ring illumination, and to a data set that concatenated the selected data from both inner- and outer-ring illumination. This latter case simulated the condition where one pair of some number, N, of different discrete sources were used for illumination at two different source-detector distances and data were collected for each of the 2N sources separately.
The results of this analysis are shown in
The ability to assess age using spectral data was tested using the NIR spectra from a multi-person study that was conducted using a laboratory-grade FTIR system similar to that shown in
The data were collected from 87 diabetic people who participated in a portion of a 17-week study. Approximately half of the people participated in the study for 6 weeks and half participated for 11 weeks. In either case, each person was measured during two separate visits per week for each week they participated in the study. During each measurement visit, multiple (3-5) optical samples were collected from the underside of their left forearm. Each optical sample consisted of 90 seconds of measurement time. A total of more than 5100 optical samples were collected on this study group. The resulting intensity spectra were log-transformed to pseudo-absorbance data and a scale function was applied to the spectra to make the spectral noise characteristics uniform. Standard outlier metrics (Mahalanobis Distance and Spectral F-Ratio) were applied to the resulting scaled absorbance data to remove outlying spectra before subsequent processing.
The scaled absorbance spectra and the corresponding ages of the subject were used in conjunction with the partial least squares (PLS) multivariate calibration algorithm to determine the age-prediction accuracy. A person-out cross validation was performed, giving the results shown in
A similar multivariate analysis was performed to determine sex prediction capability. In this case, each of the NIR spectra from the 87 subjects was assigned a reference value of either 0 or 1 based upon the sex of the person from whom the spectrum was measured. These spectral data and reference values were then processed using PLS and a subject-out cross-validation to determine sex predictions. Predicted values greater than 0.5 were assigned a value of 1 and predictions less than 0.5 were assigned a 0. The results of this analysis are given in
The ability of a spectral biometric to discriminate between live tissue and other sample types is shown in
New characteristics and advantages of the invention covered by this document have been set forth in the foregoing description. It will be understood, however, that this disclosure is, in many respects, only illustrative. Changes may be made in details, particularly in matters of shape, size, and arrangement of parts, without exceeding the scope of the invention. The scope of the invention is, of course, defined in the language in which the appended claims are expressed.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5830132 *||Feb 3, 1997||Nov 3, 1998||Robinson; Mark R.||Robust accurate non-invasive analyte monitor|
|US5867265 *||Apr 17, 1997||Feb 2, 1999||Ncr Corporation||Apparatus and method for spectroscopic product recognition and identification|
|US6041410 *||Dec 22, 1997||Mar 21, 2000||Trw Inc.||Personal identification fob|
|US6301815 *||Mar 4, 1999||Oct 16, 2001||Colt's Manufacturing Company, Inc.||Firearms and docking station system for limiting use of firearm|
|US6327376 *||Dec 3, 1998||Dec 4, 2001||U.S. Philips Corporation||Electronic apparatus comprising fingerprint sensing devices|
|US20020009213 *||Apr 11, 2001||Jan 24, 2002||Rowe Robert K.||Apparatus and method of biometric identification or verification of individuals using optical spectroscopy|
|JPH08138046A *||Title not available|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8098900 *||Mar 6, 2007||Jan 17, 2012||Honeywell International Inc.||Skin detection sensor|
|US8174394||May 8, 2012||Trutouch Technologies, Inc.||System for noninvasive determination of analytes in tissue|
|WO2013040448A1||Sep 14, 2012||Mar 21, 2013||Life Technologies Corporation||Simultaneous acquisition of biometric data and nucleic acid|
|WO2013044154A1||Sep 21, 2012||Mar 28, 2013||Life Technologies Corporation||Simultaneous aquisition of biometric data and nucleic acid|
|WO2013109612A1 *||Jan 16, 2013||Jul 25, 2013||Filmetrics, Inc.||High-lifetime broadband light source for low-power applications|
|WO2013126807A2||Feb 22, 2013||Aug 29, 2013||Life Technologies Corporation||Simultaneous acquisition of biometric data and nucleic acid|
|International Classification||G01N21/17, G06K9/00, A61B5/117, G07C9/00, A61B5/00|
|Cooperative Classification||G06K9/00, B82Y10/00, B82Y5/00, B82Y20/00, A61B5/0059, G07C9/00563, G07C9/00158, A61B5/681, A61B5/7264, A61B5/117|
|European Classification||B82Y20/00, B82Y5/00, B82Y10/00, A61B5/68B1H, G07C9/00C2D, A61B5/117, G06K9/00, G07C9/00E6|
|May 13, 2015||AS||Assignment|
Owner name: HID GLOBAL CORPORATION, TEXAS
Free format text: MERGER AND CHANGE OF NAME;ASSIGNORS:LUMIDIGM, INC.;HID GLOBAL CORPORATION;REEL/FRAME:035654/0766
Effective date: 20150401